Thermolysis of ammonium nitrate/potential donor of active chlorine compositions

Article abstract of DOI:10.1007/s10973-007-8868-z

The thermal stability of the ammonium nitrate (AN)/sodium salt of 1,3-dichlor-2,4,6-trioxo-1,3,5-triazacyclohexane (DC) composition has been studied. The factors of influence on the rates of reactions in the composition, namely, a water content, composition wetting methods, a dispersion of composition components, sample mass values, have been examined. The water presence in the composition reduces its thermal stability. The mechanism includes the partial dissociation of AN to HNO₃ and NH₃ and the hydrolysis of DC with the formation of some unstable Cl-containing compounds (chloramines, nitrogen chloride). The reaction of ammonium cation with active chlorine has been found to give rise to the explosion of the AN/DC composition. Such a situation is typical for other ammonium salt/DC compositions.

Full text of DOI:10.1007/s10973-007-8868-z

Journal of Thermal Analysis and Calorimetry, Vol. 93 (2008) 1, 301–309
THERMOLYSIS OF AMMONIUM NITRATE/POTENTIAL DONOR OF
ACTIVE CHLORINE COMPOSITIONS
Y. I. Rubtsov1, A. I. Kazakov1*, V. V. Nedelko1, Al. V. Shastin1, Tatyana S. Larikova1,
Tamara V. Sorokina1 and B. L. Korsounskii2
1Institute of Problems of Chemical Physics, Russian Academy of Sciences 1, Acad. Semenov av., 142432 Chernogolovka
Moscow region, Russian
2N. N. Semenov Institute of Chemical Physics, Russian Academy of Sciences 4, Kosygin Street, 119991 Moscow
Russian Federation
The thermal stability of the ammonium nitrate (AN)/sodium salt of 1,3-dichlor-2,4,6-trioxo-1,3,5-triazacyclohexane (DC)
composition has been studied. The factors of influence on the rates of reactions in the composition, namely, a water content,
composition wetting methods, a dispersion of composition components, sample mass values, have been examined. The water
presence in the composition reduces its thermal stability. The mechanism includes the partial dissociation of AN to HNO3 and NH3
and the hydrolysis of DC with the formation of some unstable Cl-containing compounds (chloramines, nitrogen chloride). The
reaction of ammonium cation with active chlorine has been found to give rise to the explosion of the AN/DC composition. Such a
situation is typical for other ammonium salt/DC compositions.
Keywords: ammonium nitrate, composition, explosion, kinetics, mechanism,
sodium salt of 1,3-dichlor-2,4,6-trioxo-1,3,5-triazacyclohexane, thermal decomposition
Introduction
On the other side, the studies of thermal stability of
compositions of AN with organic chlorides, that can be
donors of active chlorine in definite conditions, are
practically absent. Recently three works on physico-
chemical aspects of spontaneous explosion of AN in
the contact with donor of active chlorine have been
published [17–19] in connection with the accident in
Toulouse, France in 2001. The investigation of kinetics
and mechanism of AN reaction with these materials are
of practical and scientific interest. This reaction can
proceed in definite conditions in the mode of runaway,
and data on kinetics and mechanism of this reaction are
necessary to evaluate and predict these conditions.
AN is a well-known solid oxidant. Commonly, AN is
used as a fertilizer or as a component of industrial ex-
plosives. AN has a high thermal stability [1]. So, in
contrast with most explosives AN does not need in spe-
cial rules of manipulation. But there are some materials
forming with AN dangerous compositions of low sta-
bility. Among these materials one can find mineral ac-
ids, sulphur, some inorganic chlorides, salts of transi-
tion metals. Documents regulating the rules of AN
manipulation exclude AN contacts with these materi-
als. The list of these materials together with the analy-
sis of the main reasons of AN stability decrease one
can find in [2–5]. Among these materials, the
destabilizing effect of chloride has been studied in
great detail [6–16]. Nevertheless, chloride is some-
times added to fertilizer formulations and to permissi-
ble AN explosives. Many researchers have noted the
synergism between HNO3 and Cl– [8, 9, 11, 12], and
several claimed acid is essential for chloride catalysis,
but the acid may come from the decomposition of AN,
itself [13]. It has been suggested [7, 10] that the key
step to the destabilization is
Experimental
Materials
DC was selected as a model compound for experi-
ments with AN. DC is a potential chlorine donor
because it has two bonds N–Cl.
DC of trademark ONIACHLOR 60-G1500
(France) (with particle size distribution: 1.25 mm and
more – 7%, 0.63–1.25 mm – 80%, 0.25–0.63 mm
– 11%, and less than 0.25 mm – 2%) was used. DC par-
ticles have a well developed system of pores and capil-
laries. They can absorb large quantities of water (up to
15 mass%) without the loss of running properties.
NO+2 +Cl- ®NO2 +Cl
It has been noted that there must be an initial
excess of acid to allow formation of nitronium ion [11].
*
Author for correspondence: akazakov@icp.ac.ru
1388–6150/$20.00
Akadémiai Kiadó, Budapest, Hungary
Springer, Dordrecht, The Netherlands
© 2008 Akadémiai Kiadó, Budapest

RUBTSOV et al.
100–150 mg. When we need to carry out our experi-
ments in the presence of water the last was contained
in the separate glass. So, water could be transported
only as steam. In our experiments 100% humidity was
provided by water surplus.
Isothermal microcalorimetry
The differential automatic calorimeter DAC-1 [22]
designed and made in the Institute of Problems of
Chemical Physics, Russian Academy of Sciences was
used. It enables the heat release rate over the range of
10–5 to 1 Wt to be measured in a broad range of tem-
peratures up to 400°C. Reagents were thrown down in
a glass ampoule in two different manners: so AN was
under or above DC. The area of an ampoule internal
section was of 0.3 cm2. Actually AN and DC contacted
one to another at the points of a contact of spherical
AN granules and small DC particles. It was impossible
to determine the contact area so the heat release rates
were calculated per the unit (1 cm2) of the area of the
geometric surface between AN and DC.
(DC)
French-made industrial grade, Grand Paroisse Co.,
AN was used. It consisted of compact porous granules
(size of 0.1–0.3 cm, a granulometric density about
0.71 g sm–3). These granules strongly absorb water, so
most of water can be located on the granule surface.
When the content of surface water reaches of
3–5 mass%, the granules stick together. Further in-
crease of the surface water content results in covering
the surface of a granule by the film of a saturated wa-
ter solution of AN.
Chemically pure (Russian national standard)
ammonium chloride (AC) and potassium nitrate (PN)
were used.
Unless otherwise stated, both substances were
used as-received. Before experiments the components
were carefully dried up. In especial experiments
ground substances were used. Grinding was per-
formed with a mortar and pestle.
Mass-spectrometry
Mass-spectrometer MH1201-V (Stock Company
'Sumskoi Pribor', the Ukraine) was used to identify
gaseous reaction products. Electrons energy was of
50 eV, accelerating voltage was of 5000 V.
Moistening reagents was carried out by a direct
water addition or by sample holding in air of 90%
humidity.
IR spectrometry
Kinetic measurements were made by methods of
calorimetry, thermogravimetry and manometry. Re-
action products were identified by mass-spectrometry
and IR spectrometry.
IR spectra of condensed reaction products were
recorded on a Specord 75-IR spectrophotometer in
KBr pellets in a frequency range of 400 – 4000 cm–1.
Results and discussion
Methods
Thermogravimetry
Thermogravimetry
Thermogravimetric experiments were carried out
using electronic automatic thermobalance ATV-17M
[20] designed and made in the Institute of Problems of
Chemical Physics, Russian Academy of Sciences. A
reaction vessel – a quartz glass of a cylindrical form
(an internal diameter of 6 mm, a wall thickness of
0.1 mm) – was hung on the thin (of 0.1 mm) platinum
filament. Experiments were carried out in air at linear
heating with the rate of 4°C min–1 and at constant
temperatures.
The initial experiments with AN and DC revealed their
high thermal stability. When AN was heated linearly
(Fig. 1, curve 1) no thermal decomposition was de-
tected up to 170°C. A rapid AN decomposition was ob-
served at the temperatures above 200°C. In the process
of DC heating (curve 2) we observed a mass loss
(2.2%) in the temperature range of 80–100°C. Water
(»0.5 mass%) and some compounds of a high molecu-
lar mass (»1.7 mass%) were the products of this pro-
cess. Up to 190°C DC is thermally stable. At the tem-
peratures above 200°C DC decomposed with high
rates. In contrast to individual AN and DC their com-
position (50:50 mass%) began to release the products
at the temperatures above 90°C (Fig. 1, curve 3). At
145°C the process was sharply accelerated and at
150.5°C the composition was ignited with 20 mass%
Manometry
For manometric measurements we used Bourdon
glass manometers (reaction volume of 5–6 cm3) with
sensitive membranes [21]. Sample mass was of
302
J. Therm. Anal. Cal., 93, 2008

THERMOLYSIS OF AMMONIUM NITRATE/ACTIV CHLORINE COMPOSITIONS
experiments with two model systems: ammonium
chloride (AC)/DC and sodium nitrate (SN)/DC. When
heated linearly (4°C min–1), the composition of AC
(8 mg) and DC (8 mg) was not ignited. Thermal stabil-
ity of the AC/DC composition was not lower than that
of individual AN and DC. An addition of small quanti-
ties of water (3 mass%) into the AC/DC composition
led to an intensive heat release and foaming of the
AC/DC composition. So the AC/DC/H2O system is
very similar to the AN/DC/H2O composition. The
SN/DC and SN/DC/H2O systems were wholly inert at
heating up to 200°C. While the SN/DC/H2O composi-
tion thermolysis some DC was hydrolyzed. SN did not
react when the SN/DC/H2O composition was heated.
Fig. 1 TG curves of 1 – AN, 2 – DC and 3 – AN/DC composi-
tion in the air. Sample mass of 10.0–10.2 mg. Heating
rate of 4°C min–1
Manometry
loss. Simultaneously we observed a considerable com-
position volume increase (the final volume was 5 times
more than the initial one). It is a result of a reaction
mass swelling up.
To elucidate mechanisms of the reactions it is
important to clear up the question: either AN and DC
react only between themselves or this reaction occurs
with gaseous decomposition product participation. To
clear up the question we carried out some special
manometric experiments. AN and DC were disposed
in Bourdon manometer separately under the condition
of 100% humidity. The vessel was placed into the
thermostat at 20, 40, 60 and 80°C. A very slow form-
ation of gaseous products was noticeable at 80°C. A
special experiment when individual DC was heated at
80°C in Bourdon manometer at 100% humidity
condition proved that an observed gas evaluation in a
ternary AN/DC/H2O composition (all components
were separated) was a result of a DC hydrolysis.
Water and an environment humidity influenced
compatibility AN with DC. It is well known that AN
is a highly hydrophilic compound. When AN was in
air with 100% humidity it absorbed water, which
formed finally an AN solution in water. A reaction of
H2O with DC was a single reason of reducing AN/DC
composition stability. So if AN and DC particles are
disposed thus that, at first, water contacts with AN,
the last absorbs water and protects DC. If, at first,
water contacts with DC particles the chemical
reaction is possible and as a result we observe a low
compatibility of AN with DC. In a typical experiment
the grinding DC (53 mg) was placed at the bottom of
the glass cylindric vessel (an internal diameter of
5 mm), then a water drop (9 mg) was placed on the
DC layer and finally the AN layer (52 mg) was
disposed on the top of the composition (the water
content was 8.6 mass%). The chemical reaction in this
composition started immediately at 18°C. The initial
rate of gas formation at 18°C was 0.42 mL g–1 min–1.
Initial rates of the process at 30 and 40°C were 1.1
and 2.5 mL g–1 min–1. The value of 'effective' acti-
vation energy of the reaction in the AN/DC/H2O
composition was equal to 76.1 kJ mol–1.
As it is shown in Fig. 2 at 140°C the components
of the AN/DC composition (50:50 mass%) react with
moderate rates (Fig. 2, curve 1). The total time of the
process was about 40 min. AN and DC grinding gave
rise to a sharp reaction rate increase. The AN/DC
composition consisted of grinding components was
ignited at 140°C after 25 s (Fig. 2, curve 2). It is
interesting to note that a mass loss (»20%) was the
same as in the previous experiment (Fig. 1, curve 3).
Moistening composition and grinding its components
led to the process acceleration. Even without grinding
the composition of dry DC and moistened AN
(10.2 mass% H2O) began to decompose at 47°C.
To understand the mechanisms of chemical reac-
tions in the AN/DC composition we carried out some
Fig. 2 Kinetics of chemical reaction in AN/DC composition
(50:50 mass%) in air at 140°C . 1 – as-received compo-
nents, 2 – ground components. Sample mass of 20 mg
J. Therm. Anal. Cal., 93, 2008
303

RUBTSOV et al.
(~2–4·10–2 mW cm–2) connected with vaporization of
water, which was contained in the initial samples, was
observed. The possible heat release rate due to interac-
tion of AN and DC must be lower than this level. The
very small heat absorption rate was also observed
when 3.7 mass% of water was added to DC. When
7 mass% of water was added to DC the initial peak of a
heat release rate of 22 mW sm–2 was observed. Then
the heat release rate decreased so fast that in 10 min it
was twenty times smaller of the initial value. In 3 h the
heat release was about 5.3 J sm–2. The measured initial
rate of DC hydrolysis was about 1.5 mW sm–2, after
6 min the rate was 10 times smaller, and after 25 min it
approached to zero. The experiment results with direct
water additions are listed in Table 1. The data on sam-
ple wetting by their exposure in air of a definite humid-
ity are listed in Table 2.
Fig. 3 Mass-spectrum of volatile at 25°C products of AN/DC
(50:50 mass%) composition thermolysis (116°C, 30 min)
When AN surface wetted, the oscillations of the
rate in the course of the reaction were practically not
observed as AN and DC granules were strongly
adhered one to another due to the presence of signi-
ficant amounts of water on the AN surface. In these
conditions, the reaction proceeded on the practically
constant contacts. The rate decrease determined by
decrease of reactant contents on the contacts due to
the replacement of some DC by its hydrolysis
product, which did not contain chlorine.
Mass-spectrometry
Mass spectrum of gaseous products of the AN/DC
composition thermolysis is demonstrated in Fig. 3. The
thermolysis at 116°C for 30 min was accompanied by
formation of N2, N2O and H2O (typical products of AN
thermal decomposition) and also HCl, Cl2 and very
small quantities of Cl–CºN.
The heat release rate did not depend on masses
of components (experiments No. 3–8), and it was
proportional to the area of a geometrical surface of a
contact of components. In ampoules with a different
inner diameter (experiments No. 56, 57) the heat
release rate decreased in 1.2 times at decreasing a
sectional area of the ampoule in 1.1 times. The heat
release rate also increased essentially at preliminary
intermixing AN and DC samples in the bottle (experi-
ments No. 8 and 9, 52 and 53). The heat release rate
for preliminary intermixed components (mass of
0.1 g) was 3 times higher than that in the layer system
without intermixing. The greater a zone of intermixed
components was, the higher the heat release rate was.
The process spread all over the volume of intermixed
AN and DC.
IR spectrometry
According to IR spectroscopy data isocyanuric acid
and its derivatives were formed as condensed products
of the AC/DC/H2O composition thermolysis. In the
case of the AN/DC/H2O composition these products
were not observed. It is important to note that Cl– ions
(6 mass%) were detected among condensed products
of the AC/DC/H2O composition ignition.
Isothermal microcalorimetry
For the moistened AN/DC compositions the chaotical
oscillations of the reaction rate in the course of
reaction were observed. Probably, the reason of this
phenomenon was a change of the real area of surface
of AN contacts with DC in the course of the reaction.
The former contacts lost and the new ones formed.
The rate changes observed in the case of moistened
DC may be also the result of different rates of a water
transport in the DC capillary system. To obtain more
correct data on the heat release rates we found
average values of experimental reaction rates. For this
purpose a rate curve was integrated in a certain time
interval. Then the calculated heat was divided by the
time of integration.
In Fig. 4 the dependence of the average heat
release rate (W cm–2) on water content in the AN/DC
composition is demonstrated. DC was moistened in
this experiment. Up to 4.1 mass% H2O content in DC
the process occurred with heat absorption because of
an endothermal H2O evaporation prevailed over an
exothermal hydrolysis and oxidation. When water
content in DC was of 4.1–6.4 mass% the exothermal
processes predominated and we observed the heat
release. The heat release rates did not depend on the
water content in DC above 6.4 mass%. Probably,
most part of water in DC disposed inside DC particles
For a layering composition of 0.1 g AN and
0.1 g DC without humidification of components at
25°C the very small heat absorption rate
304
J. Therm. Anal. Cal., 93, 2008

THERMOLYSIS OF AMMONIUM NITRATE/ACTIV CHLORINE COMPOSITIONS
Table 1 Results of microcalorimetric experiments at a direct water addition to the samples
Component
Q/J cm–2
mAN/g
mDC/g
(dQ/dt)t=0 /mW cm–2
No.
T/°C
below
2
3
25.0
64.7
64.7
64.7
64.7
64.7
64.7
64.7
53.3
53.3
53.3
53.3
47.5
47.5
41.0
41.0
41.0
37.0
37.0
0.0
0.020
0.106+4.1% H2O
0.096+6.7% H2O
0.103+5.6% H2O
0.101+6.8% H2O
0.101+12.3% H2O
0.102+14% H2O
0.103+12.7% H2O
0.102+13.7% H2O
0.104+14.5% H2O
0.102+13% H2O
0.100+13.8% H2O
0.105
DC
1.5
28
0.96 in 38.5 min
59 in 160 min
128 in 275 min
109 in 155 min
91 in 165 min
82.6 in 182.5 min
100 in 200 min
60 in 43.5 min
22 in 50 min
DC
4
0.046
AN
31
5
0.041
DC
33
6
0.114
DC
35
7
0.042
AN
39
8
0.252
DC
27
9*
10
11
12
15
16
18
19
20**
21**
22
24**
0.103
DC
83
0.105
AN
13.6
14.2
6.7
18
0.099
DC
75 in 262.5 min
64 in 205 min
27 in 200 min
70 in 367.5 min
105 in 120 min
140 in 480 min
4.7 in 420 min
150 in 490 min
44 in 270 min
280 in 970 min
0.102
AN; DC grinded
0.101+7.1% H2O
0.103
AN
DC
AN
DC
–
0.102+12.7% H2O
0.106
8.6
52
0.100+10.0% H2O
0.041
0.097+12.4% H2O
0.101+12.5% H2O
0.101+8.0% H2O
0.102+15.1% H2O
0.102
6.1
1.9
5
0.0
0.018
DC
AN
AN
0.105
4.4
19.4
0.100+12.6%H2O
*components were intermixed in the bottle; **ampoule sealed
Fig. 4 The dependence of the initial heat release rate on the
Fig. 5 The dependence of the initial heat release rate on the
water content in AN/DC (moistened) composition
water content in AN(moistened)/DC composition
but the water content on the DC particle surface was
close to a constant value.
moistened the heat release rates were significantly
greater (5–10 times). In contrast to DC, most part of
water disposed on the AN granule surface. The greater
rates of the exothermal processes in the case of the
moistened AN could be explained by the greater
quantities of water which takes part in a DC hydrolysis
and by the greater concentration of products of AN
dissociation equilibrium (HNO3 and NH3) at the places
of AN and DC contacts. From results of experiments
Another situation was observed when the AN/DC
composition contained a preliminary moistened AN
(Fig. 5). The exothermal processes dominated at any
water content in AN. The comparison of two compo-
sitions: AN (moistened)/DC and AN/DC (moistened)
reveals that the heat release rates were strongly
dependent on the water content in AN. When AN was
J. Therm. Anal. Cal., 93, 2008
305

RUBTSOV et al.
Table 2 Results of microcalorimetric experiments at introducing water by the exposure of the samples in wet air
Component
below
(dQ/dt)t=0 /
mW cm–2
Q/J cm–2
mAN/g
mDC/g
No.
T/°C
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53*
54**
64.2
64.7
64.7
64.7
74.0
74.0
64.0
74.0
74.0
54.5
60.6
54.5
60.6
60.2
45.3
50.3
45.3
49.3
49.3
45.3
45.3
45.3
39.3
33.6
33.6
25.4
25.4
25.4
0.1037
0.09915
0.0892+4.1% H2O
0.0851+11.1% H2O
0.971
AN
AN
AN
AN
AN
AN
AN
AN
AN
DC
DC
DC
DC
AN
AN
AN
AN
AN
AN
AN
AN
AN
AN
AN
AN
AN
AN
AN
–0.52
28
–7.6 in 450 min
45.7 in 114 min
24.1 in 96 min
102 in 150 min
32.4 in 231 min
22 in 48 min
0.0974+5.1% H2O
0.0943+8.2% H2O
0.10235
55.4
178
100
64
0.098
0.0901+6.4 % H2O
0.0965+9.7% H2O
0.10475
0.102
0.0911+2.69% H2O
0.0983
6.8
72
2.7 in 15 min
0.0883+11.1% H2O
0.0992+4.1% H2O
0.10055
18 in 10.2 min
–5.0 in 228 min
13.6 in 87.6 min
11.0 in 60 min
72.8 in 198 min
76 in 150 min
0.1063
–3.8
7.69
15.0
22
0.0975+5.1% H2O
0.0923+5.1% H2O
0.1004
0.103
0.0875+11.1% H2O
0.09247+11.1% H2O
0.1045
0.09425
33
0.0898+5.1% H2O
0.0911+5.1% H2O
0.0943+5.1% H2O
0.0931+12.37% H2O
0.1022
32
10.2 in 25.2 min
14.6 in 100.2 min
12.0 in 100.2 min
117 in 199.8 min
45.7 in 207.6 min
53.2 in 252 min
64.2 in 225 min
6.2 in 84 min
0.105
22
0.10115
168
246
13
0.103
0.0826+11.1% H2O
0.064+11.1% H2O
0.099
0.10065
5
0.0859+8.2% H2O
0.1022+2.69% H2O
0.0945+10.0% H2O
0.0974+5.1% H2O
0.0879+5.1% H2O
0.0957+5.1% H2O
0.0938+5.1% H2O
0.0873+5.1% H2O
0.10175
85
0.1006
5.6
111
9.8
18
0.10265
88 in 267 min
0.10635
10.6 in 300 min
14.0 in 225 min
53.7 in 420 min
14.3 in 174 min
15.5 in 186 min
145 in 297.6 min
0.1089
0.0910+13.0% H2O
0.0999
5.6
9.6
26.9
342
0.1046
0.1042
5.75% solution
of DC in H2O
0.0157+0.2580 g H2O
55
25.4
0.0957
2.1
11.9 in 360 min
56
57
58
25.4
25.4
74
0.0882+5.1% H2O
0.0948+5.1% H2O
0.0932
0.1053
0.9965
0.10065
AN
AN
AN
7.8
7.3
–3
7.1 in 402 min
8.4 in 276 min
*components were intermixed in the bottle; **0.1483 g of water was poured into the ampoule with the composition
No. 50 and 51 it has been seen that humidifying both
components even decreases the heat release rate that in
contrast with using AN humidified only. It is probable
that the lower water content on the DC surface in
contrast to AN can not change the total water content
in the places of the AN and DC contacts. The
vaporization of water from DC humidified reduces the
total heat release rate. These results demonstrate that
AN humidifying is most dangerous. Thus, calorimetric
experiments supported the important role of water in
the reaction of AN with DC.
In the experiment with an open ampoule the
process stopped because of evaporation and some part
of AN and DC did not react. When the ampoule was
sealed (experiments No. 21 and 24) all the gaseous
reaction products and water were retained in the
reaction zone so we observed a process self-accel-
eration and its completion. The observed self-accel-
eration of the process was a result of a water content
increase as the reaction of AN with DC accompanied
by ammonium ion oxidation with water formation.
The total heat of the process was 2.2 kJ g–1 AN. This
306
J. Therm. Anal. Cal., 93, 2008

THERMOLYSIS OF AMMONIUM NITRATE/ACTIV CHLORINE COMPOSITIONS
value was higher than the total heat of AN decom-
position. It means that stoichiometry of the process
differs from stoichiometry of the AN decomposition
and that DC was active in oxidation reactions. When
one of the components was moistened and the reac-
tion composition was placed in the sealed ampoule,
water in quantities sufficient to hydrolyze DC was
disposed all over the reaction composition. AN was
accessible to gaseous products of a DC hydrolysis so
an AN oxidation by these products took place all over
the volume. As a result the reaction heat calculated
per the unit of ampoule section area was proportional
to AN mass.
The dependence of average values of the heat
release rate for an initial reaction stage in the case of
the AN/DC (preliminary moistened) layer composi-
tion with water content of 12–14 mass% on the
temperature (interval 37.0–64.7°C) was described by
the equation
Fig. 7 Arrhenius dependence of the initial heat release rate
(average value) for AN (preliminary moistened)/DC
composition
(dQ/dt)t=0=8.5·108exp(–67000/RT) W cm–2
This dependence is demonstrated in Fig. 6.
The heat release for the AN (moistened)/DC
composition is the process of the first kinetic order
(up to 70–80% conversion) in the temperature range
of 25.5–66.0°C. Arrhenius dependence of the initial
heat release rate is shown in Fig. 7. This dependence
is described by Arrhenius equation
To understand correctly kinetic regularities of re-
action it is necessary to define the stage of the complex
process which determines its rate. Considering an easy
hydrolysis of DC by water with the release of chlorine
and the necessity of water for interaction AN and DC,
one can propose that namely the hydrolysis of DC de-
termines the rate of the total process. However, this re-
action is rather slow. At the ambient temperature the
chlorine release from the aqueous solution of DC pro-
ceeds for several days. At the same time pouring ap-
proximately 42% of water (counting per the total mass)
in the calorimetric ampoule with the dry layering
AN/DC composition resulted in the sharp increase of
the heat release rate from zero value up to 340 W sm–2
(experiment No. 54). The very intensive release of
chlorine and other gases is also observed directly at
filling the ampoule by water after carrying out an ex-
periment to dissolve AN and DC stayed. It was possi-
ble to assume that AN increases the DC hydrolysis rate
very sharply, and chlorine releasing interacting with
AN. AN always contains some equilibrium quantities
of ammonia and nitric acid, which can cause an
acid-base catalysis of a hydrolysis of N–Cl bond in
DC. To check this supposition, the experiment at 25°C
(No. 55) with 5.75% aqueous solution of DC placed in
the microcalorimeter ampoule was conducted using the
microcalorimeter. After some time of the record of the
heat release rate at DC hydrolysis AN sample was
dropped in this solution. The rate of a hydrolysis was
increased from value of 7 mW g–1 DC in aqueous solu-
tion up to 35 mW g–1 DC in the presence of AN due to
its catalytic influence. These factors in connection with
high rates of water absorbing by AN granules are
(dQ/dt)t=0=1.35·107exp(–54000/RT) W cm–2
The lower value of activation energy for the
composition with preliminary moistened AN in
comparison with that of the AN/DC (preliminary
moistened) composition is probably a result of whole
water disposition on the AN granule surface.
Fig. 6 Arrhenius dependence of the initial heat release rate
(average value) for AN/DC (preliminary moistened)
composition
J. Therm. Anal. Cal., 93, 2008
307

RUBTSOV et al.
responsible for a fast process in the AN/DC compo-
and DC contact surface decreases sharply and the
whole process also ceases.
sition when AN was moistened.
The dry AN/DC composition was highly ther-
mally stable and safe. But it is important to note that
this composition decomposes at the temperatures
above 100°C in contrast to individual AN and DC,
which are stable up to 190–200°C. Moistening compo-
sition (high atmosphere humidity or water addition)
sharply decreased the compatibility of AN with DC in
this composition. The action of water is rather com-
plex. At first, dissolving AN and DC in water promotes
the interaction between reagents because of homogeni-
zation of the reaction composition. At second, water
takes part in some reactions. The experimental data
prove that water hydrolyzes DC:
The results of calorimetric experiments show
clearly that AN moistening accelerates the reaction of
AN with DC more than DC moistening. But this
statement does not contradict a main role of
reaction (1). Being highly hydrophilic AN acts as
water container, which moistens DC.
Undoubtedly, the number of chemical processes
in the AN/DC/H2O composition is not limited by re-
actions (1)–(7). It is known that chloride-ions can ac-
celerate the AN decomposition at temperatures higher
m.p. [8]. But at temperatures under study because of a
high AN thermal stability [23] the role of this process
may be not so important.
The most relevant factors to enhance possibility
of thermal explosion in the system studied are to
humidify AN on the surface up to the general water
content of 5% and more and to intermix composition
components even partially in the thin layer. The
factors that counteract the thermal explosion are an
easy vaporization of water at high temperatures and
rather a low activation energy for the heat release rate.
When the separate layers system does not possess of
excess water the reaction practically does not
proceed; also the thermal spontaneous combustion is
impossible.
>N–Cl+H2O®>N–H+HOCl
The next reactions are possible:
(1)
2HOCl®H2O+Cl2O
Cl2O®Cl2+0.5O2
(2)
(3)
Probably small quantities of free Cl and O are
formed as intermediate reaction products and a
composition activity in different oxidations increases
strongly. The next very important reaction is the
reaction of chlorine and probably another chlorine-
containing DC hydrolysis products with ammonium
ion and ammonia (a product of AN dissociation):
Conclusions
2NH3+3Cl2®N2+6HCl
(4)
This reaction occurs via intermediate mono-
chloramine, dichloramine and nitrogen trichloride
NCl3. The important role of reaction (4) has been
proved by the experimental fact that both AN and AC
react with DC but SN does not. So ammonium ion
(but not nitrate ion) is a reaction species interacting
with DC. HCl is active in a reaction with HNO3
forming from AN:
The interaction of AN and DC in the not intermixed
composition at temperatures up to 65°C proceeds only
in the thin layer on the contacts of components. The heat
release rate depends on temperature and the interfacial
area of components and does not depend on the masses
of components. Introducing water in AN or DC
increases sharply the heat release rate, AN humidifying
influencing much more strongly. Intermixing compo-
sition components results in the essential increase of the
rate as the reaction proceeds in the total volume of
components intermixed. At humidifying DC the process
proceeds with large oscillations of the heat release rate,
at introducing water in AN these oscillations are small
or not observed at all. There is not the heat release in the
dry composition up to temperature 65°C, in the open
ampoule even the very small heat absorption is
observed. The kinetic regularities of the heat release, the
dependences of its rate on the temperature and the water
content have been obtained. These data can be used to
calculate any thermal processes in the system studied.
HNO3+3HCl®2H2O+NOCl+Cl2
2NOCl®2NO+Cl2
(5)
(6)
The products of reactions (5) and (6) are capable
to oxidize ammonia also:
NOCl+NH3®N2+H2O+HCl
(7)
In these reactions we observe N2 and NO
formation (but no N2O, which is the main AN
decomposition product). N2 and NO formation
instead of N2O increases the reaction heat. Moreover,
NO can oxidize DC forming CO2 with an additional
heat release. So, chlorine consumes in reaction (4)
and forms in reactions (5) and (6), and evaporates
partially. So when a DC hydrolysis ceases because of
water evaporation, the chlorine content on the AN
308
J. Therm. Anal. Cal., 93, 2008

THERMOLYSIS OF AMMONIUM NITRATE/ACTIV CHLORINE COMPOSITIONS
12 H. L. Saunders, J. Chem. Soc., 121 (1922) 698.
References
13 H. Tramm and H. Velde, Angew. Chem., 47 (1934) 782.
14 B. J. Wood and H. Wise, J. Chem. Phys., 23/24 (1955) 693.
1 G. B. Manelis, G. M. Nazin, Yu. I. Rubtsov and
V. A. Strunin, Thermal Decomposition and Combustion of
Explosives and Propellants, Taylor and Francis Group,
London and New York 2003, p. 175.
15 W. A. Rosser, S. H. Inami and H. Wise, Trans. Faraday Soc.,
67 (1963) 1753.
16 K. S. Barclay and J. M. Crewe, J. Appl. Chem.,
17 (1967) 21.
2 M. A. Budnikov, N. A. Levkovich, I. B. Bystrov,
V. F. Cirotinskii and B. I. Shekhter, Vzryvchatye
Veshestva and Poroha (Explosives and Powders),
Moscow, Publishing House ‘Oborongiz’ 1955, p. 364
(in Russian).
17 Y. Guengant, P. Della Pieta and M. Dervaux, Proceedings
of 28th Annual General Meeting of Federation of European
Explosives Manufacturers (FEEM). Copengagen,
Denmark, (2004) p. 28.
18 C. M. Badeen, Q. S. M. Kwok, M. C. R. Vachon,
R. Turcotte and D. E. G. Jones, J. Therm. Anal. Cal.,
81 (2005) 225.
3 A. P. Glazkova, Kataliz Goreniya Vzryvchatyh Veshestv
(Catalysis of Explosives Combustion), Publishing House
‘Nauka’, Moscow 1976, p. 263 (in Russian).
4 Z. G. Pozdnaykov and B. D. Rossi, Spravochnik po
Promyshlennym Vzryvchatym Veshestvam (Reference
Book on Industrial Explosives), Publishing House
‘Nedra’, Moscow 1977, p. 255 (in Russian).
5 L. V. Dubnov, I. S. Baharevich and A. I. Romashov,
Promyshlennye Vzryvchatye Veshestva (Industrial
Explosives), Publishing House ‘Nedra’, Moscow 1973,
p. 319 (in Russian).
19 B. S. Ermolaev, A. A. Sulimov, A. A. Borisov,
B. L. Korsunskii, M. K. Sukoyan, V. A. Foteenkov,
P. V. Komissarov, G. N. Sokolov, V. V. Nedelko,
A. V. Shastin, A. I. Kazakov, Y. I. Pubtsov, A.-N. Prel,
B. A. Khasainov and P. Vidal, Khim. Fiz., 26 (2007) 50.
20 L. N. Gal’perin and Yu. R. Kolesov, Izmerit. Tekhnika
[Measuring Technique], 4 (1981) 23.
21 K. K. Andreev, Termicheskoe Razlojenie i Gorenie
Vzryvchatyh Veshestv (Thermal Decomposition and
Combustion of Explosives), Publishing House ‘Nauka’,
Moscow 1966, p. 339 (in Russian).
6 A. G. Keenan, J. Am. Chem. Soc., 77 (1955) 1379.
7 A. G. Keenan, K. Notz and N. B. Franco, J. Am. Chem. Soc.,
91 (1969) 3168.
22 L. N. Gal’perin, Yu. R. Kolesov and L. B. Mashkinov,
Proceedings of 6th All-Union Conf. on Calorimetry,
Tbilisi, (1973) 539 (in Russian).
8 Y. I. Rubtsov, I. I.Strizhevskii and A. I. Kazakov,
Zh. Prikl. Khim., 62 (1989) 2417.
9 Y. I. Rubtsov, I. I. Strizhevskii and A. I. Kazakov,
Zh. Prikl. Khim., 62 (1989) 2169.
23 Y. I. Rubtsov, A. I. Kazakov and L. P. Andrienko,
Zh. Prikl. Khim., 60 (1987) 3.
10 A. G. Keenan and B. Dimitriades, J. Chem. Phys.,
37 (1962) 1583.
11 C. I. Colvin, P. W. Fearnow and A. G. Keenan,
J. Inorg. Chem., 4 (1965) 173.
DOI: 10.1007/s10973-007-8868-z
J. Therm. Anal. Cal., 93, 2008
309